Loading [a11y]/accessibility-menu.js
ReLDA for micro-blog topic mining based on relationship structure | IEEE Conference Publication | IEEE Xplore

ReLDA for micro-blog topic mining based on relationship structure


Abstract:

Micro-blog is a sort of open social network platform. The massive real-time messages that Twitter users publish through the micro-blog platform every day are called tweet...Show More

Abstract:

Micro-blog is a sort of open social network platform. The massive real-time messages that Twitter users publish through the micro-blog platform every day are called tweets, which are characterized with instantaneity and subjectivity and imply some topic information that contain the users' daily activities and hot news. Because of the short length and non-prominent topics of tweets, the traditional mining and clustering methods are not very effective. In order to solve the difficulty of topics mining, we try to find the relationship structure by the dependent of tweets, and propose reLDA (the abbreviation of relationship LDA) model based on the relationship structure. In particular, we build the tree structure for tweets, then deduce and improve the process of Gibbs Sampling by this model, at last we carry out the topic mining of tweets that captured by Twitter API, the result of experiment proves that our model is effective.
Date of Conference: 13-15 August 2016
Date Added to IEEE Xplore: 24 October 2016
ISBN Information:
Conference Location: Changsha, China

References

References is not available for this document.